Article
Orthopedics
Zachary C. Lum
Summary: Recent advances in neural networks, deep learning, and AI have led to the development of a new AI model called ChatGPT that uses large language models. However, implementing knowledge with this model remains a challenge.
CLINICAL ORTHOPAEDICS AND RELATED RESEARCH
(2023)
Article
Medicine, General & Internal
Adem Gencer, Suphi Aydin
Summary: This study tested the performance of ChatGPT on Turkish-language thoracic surgery exam questions and found that it achieved good results, even surpassing students. This indicates that ChatGPT can also be successful in specific areas of medicine.
AMERICAN JOURNAL OF THE MEDICAL SCIENCES
(2023)
Article
Medicine, General & Internal
Rory Cuthbert, Ashley Simpson
Summary: Purpose Chat Generative Pre-trained Transformer (ChatGPT) is a large language AI model that generates contextually relevant text in response to questioning. This study assessed whether ChatGPT could pass the FRCS examination in Trauma and Orthopaedic Surgery. The results showed that ChatGPT currently lacks the higher-order judgement and multilogical thinking required to pass the examination and fails to recognize its own limitations.
POSTGRADUATE MEDICAL JOURNAL
(2023)
Article
Orthopedics
Janina Kaarre, Robert Feldt, Laura E. Keeling, Sahil Dadoo, Balint Zsidai, Jonathan D. Hughes, Kristian Samuelsson, Volker Musahl
Summary: This study investigated the potential use of large language models (LLMs) in orthopaedics by presenting queries related to ACL surgery to a generative pre-trained transformer (ChatGPT). The findings suggest that ChatGPT can generate correct responses in approximately 65% of the cases related to ACL surgery. While LLMs offer potential as a supplementary tool for acquiring orthopaedic knowledge, they cannot replace the expertise of orthopaedic sports medicine surgeons in diagnostic and treatment planning endeavours.
KNEE SURGERY SPORTS TRAUMATOLOGY ARTHROSCOPY
(2023)
Review
Orthopedics
Hai Le, Joseph B. Wick, Brian M. Haus, George S. M. Dyer
Summary: The Orthopaedic In-Training Examination (OITE) is a standardized national test administered annually to orthopaedic residents by the American Academy of Orthopaedic Surgeons since 1963. It consists of 275 multiple-choice questions covering 11 domains of orthopaedic knowledge and is considered predictive of success in residency and the American Board of Orthopaedic Surgery part I examination. Residents are advised to start preparation early, focus on comprehension over memorization, and avoid burnout.
JOURNAL OF THE AMERICAN ACADEMY OF ORTHOPAEDIC SURGEONS
(2021)
Article
Radiology, Nuclear Medicine & Medical Imaging
Rajesh Bhayana, Satheesh Krishna, Robert R. Bleakney
Summary: ChatGPT is a powerful AI language model with potential in medical practice and education, but its performance in radiology is uncertain.
Article
Computer Science, Information Systems
Yaa Kumah-Crystal, Scott Mankowitz, Peter Embi, Christoph U. Lehmann
Summary: This study evaluated the performance of ChatGPT on the Clinical Informatics Board Examination and discussed the implications of large language models (LLMs) for board certification and maintenance. ChatGPT was tested using 260 multiple-choice questions and answered 190 (74%) of the eligible questions correctly. The findings raise concerns about potential misuse and validity of knowledge assessment exams due to the accuracy of ChatGPT in answering multiple-choice questions.
JOURNAL OF THE AMERICAN MEDICAL INFORMATICS ASSOCIATION
(2023)
Article
Medicine, General & Internal
Ying-Mei Wang, Hung-Wei Shen, Tzeng-Ji Chen
Summary: ChatGPT achieved a correctness rate of 54.4% and 56.9% in the Chinese and English questions for the Taiwanese Pharmacist Licensing Examination. Only the pharmacology and pharmacochemistry sections passed the examination. Although ChatGPT did not pass the examination, it can be quickly improved through deep learning. This study highlights the need to use a variety of evaluation methods and incorporate AI technology into pharmacy education.
JOURNAL OF THE CHINESE MEDICAL ASSOCIATION
(2023)
Article
Surgery
Pooja Humar, Malke Asaad, Fuat Baris Bengur, Vu Nguyen
Summary: ChatGPT, an artificial intelligence language model developed by OpenAI in 2022, was evaluated for its performance on the Plastic Surgery In-Service Examination. Results showed that ChatGPT performed at the level of a first-year resident but poorly compared to more advanced residents. Further research is needed to assess its efficacy despite its potential uses in healthcare and medical education.
AESTHETIC SURGERY JOURNAL
(2023)
Article
Obstetrics & Gynecology
Sarah W. Li, Matthew W. Kemp, Susan J. S. Logan, Pooja Sharma Dimri, Navkaran Singh, Citra N. Z. Mattar, Pradip Dashraath, Harshaana Ramlal, Aniza P. Mahyuddin, Suren Kanayan, Sean W. D. Carter, Serene P. T. Thain, Erin L. Fee, Sebastian E. Illanes, Mahesh A. Choolani
Summary: This study investigated the performance of ChatGPT in engaging with healthcare systems and completing a mock objective structured clinical examination. The results showed that ChatGPT outperformed human candidates in several knowledge areas and generated accurate and contextually relevant answers.
AMERICAN JOURNAL OF OBSTETRICS AND GYNECOLOGY
(2023)
Article
Education, Scientific Disciplines
Andrew Mihalache, Ryan S. Huang, Marko M. Popovic, Rajeev H. Muni
Summary: ChatGPT-4 demonstrated high accuracy in responding to practice questions for USMLE examinations, with explanations provided for all questions. The AI chatbot spent around 30 seconds on average per question, with similar lengths for correctly and incorrectly answered questions, but significantly shorter lengths for correct responses compared to incorrect responses.
Article
Medicine, General & Internal
Wenting Tong, Yongfu Guan, Jinping Chen, Xixuan Huang, Yuting Zhong, Changrong Zhang, Hui Zhang
Summary: This article examines the application and limitations of AI technology in Chinese and English settings, and evaluates ChatGPT's performance in the 2022 National Medical Licensing Examination (NMLE) in China through an experiment. The study found that ChatGPT had a correct response rate of 81.25% for Chinese questions and 86.25% for English questions. However, language bias remains a significant challenge for AI language models.
FRONTIERS IN MEDICINE
(2023)
Article
Education & Educational Research
Austin Pack, Jeffrey Maloney
Summary: Progress in NLP and AI has made tools like OpenAI's GPT-3.5 more accessible for non-experts. While discussions mainly focus on AI's impact on language education, this paper highlights the potential of generative AI for language education researchers. It demonstrates how ChatGPT can assist researchers in various ways and raises ethical concerns for using AI in research.
Article
Education & Educational Research
Yun Dai, Sichen Lai, Cher Ping Lim, Ang Liu
Summary: This study explores the impact of advanced AI on research supervision of postgraduate students. A qualitative approach was used to examine the practices and perspectives of 20 postgraduate research students in Australia. The study finds that ChatGPT can accelerate research progress, enhance research quality, improve scholarly development and critical thinking, increase student confidence and autonomy, and deepen supervisory relationships. The study also suggests a shift in the roles and responsibilities of supervisors and students, with AI acting as epistemic tools to enhance the supervisory process. Ethical implications of AI-enabled support are also considered.
AUSTRALASIAN JOURNAL OF EDUCATIONAL TECHNOLOGY
(2023)
Article
Clinical Neurology
Gage A. Guerra, Hayden Hofmann, Sina Sobhani, Grady Hofmann, David Gomez, Daniel Soroudi, Benjamin S. Hopkins, Jonathan Dallas, Dhiraj J. Pangal, Stephanie Cheok, Vincent N. Nguyen, William J. Mack, Gabriel Zada
Summary: This study examines the competence of GPT-4, an updated language model, on neurosurgical board-style questions. The results show that GPT-4 outperforms medical students and residents, suggesting its potential in medical education and clinical decision-making.
WORLD NEUROSURGERY
(2023)